Managing Crowdsourced Communities

Crowdsourcing's future depends on change management

David Butler interviews Frank Piller via telephone on May 13, 2007.

Frank Piller is a chair professor of management at the Technology & Innovation Management Group of RWTH Aachen University, Germany, one of Europe’s leading institutes of technology. He is also a founding faculty member of the MIT Smart Customization Group at the Massachusetts Institute of Technology, USA. Before entering his recent position in Aachen in March 2007, he worked at the MIT Sloan School of Management (2004-2007) and has been an associate professor of management at TUM Business School, Technische Universitaet Muenchen (1999-2004).

On May 13th, I had the pleasure of interviewing Dr. Frank Piller on crowdsourcing for the Assignment Zero project. We spent an hour discussing the history, current examples and future implications of the crowdsourcing phenomena as described by Jeff Howe. Dr Piller's expertise is on how firms can co-create value with their customers.

David Butler: What were some of the factors that led to your interest in this field?

Dr. Frank Piller: My background has been research in mass customization, and customizing consumer goods and high efficiency. While my Ph.D is in Operations Management, I later got more and more interested in front end issues. Here, the larger changes occur when companies invite their consumers to co-design or configure a product. The next step in my research was to investigate how firms include customers in the innovation process – the next logical step after co-design in a mass customization configurator. Very often what is happening if you just configure a product, you also learn more about a product. As a consumer you may get an idea on how to improve a product. The next step was: If this is a way for companies to get input, how can they facilitate this input? During the years I spent at MIT, I looked into ways that companies can provide tools and processes so that external actors and mostly customers and users can provide this input. Finally I found that the larger topic that really interested me as a management professor is the new way to organize the share of labor. This is a very basic economic principle. That was around the time when Jeff came up with his crowdsourcing article and provided somehow a larger framework.

Q: So the crowdsourcing concept acts as an umbrella framework for several different roots?

A: Yes. Yale professor Yoachi Benkler has described this economic principle perfectly in his book, “The Wealth of Networks”: Instead of either buying a product or service on the market or assigning a task within your hierarchy (to one of your employees), you ask an open network of potential contributors to provide this task. Jeff’s crowdsourcing article made this concept much more feasible and applicable. I found that mass customization also relates to Jeff Howe’s definition. A company places its configurator openly on the Internet and then consumers can select if they take this extra effort to design a product or if they just buy a standard product. This is the open call for participation in Jeff’s definition.

Q: So the customer has a choice?

A: Exactly, no one has to buy a custom product and spend the time to co-design it.

Q: I especially liked your reference to Learning Relationships in your article, "The Personal Touch" in SAP Info.

A: Customizing and co-designing first is a way to increase efficiency, to get rid of inventory or reduce the forecasting risk. The term "Learning Relationships" (originally coined by consultants Don Peppers & Martha Rogers) describes another benefit, a new way for companies beyond traditional market research or traditional CRM to learn about customers to provide additional value. Companies recognize that integrating customers into value creation is a new way to tap into external input. The result is not a one way street but a learning cycle.

Q: So, for example, Amazon knowing what kinds of books I buy and making recommendations, they have my information……

A: Yes, absolutely. The longer you are a client with Amazon not just buying books but also providing reviews and recommendations, the better the feedback is. To get me right: this is not real crowdsourcing but a type of CRM in a more clever way. But the line blurs. You as a consumer do a lot of voluntary work for Amazon, more than at a normal bookstore. As a result, they can serve you better. But it also becomes easy for you to stay in business with them. This is where Amazon has advanced approaches that make it unique.

Q: Does this take us to crowdsourcing 2.0?

A: Yes, I think this is one important element from the firms’ perspective. Amazon makes it easy for you to shop with them and you don’t want to change providers. In the future, also, more typical crowdsourcing companies like Threadless need to improve their user interfaces to make it easy to co-create and design for users with them.

At the moment, firms like Threadless do not have the extra services of an Amazon. Now they show you T-shirt designs, all of them, where they could in the future base what designs they show you on what they know about your design preferences and previous decisions. The companies that will succeed in a world where more and more companies compete on the input of their users are those that have made it easy to interact with them - and create the most pleasurable crowdsourcing experience.

Q: I note your reference to how real estate sales of condominiums get a certain number of commitments before they build, which is similar to how Threadless determines its investment into clothing.

A: This is a traditional way in condominiums but this is a new model where Threadless applies this to cheap consumer goods like T-shirts.

Q: Coming back to Threadless and mass customization. This reminds me of a late 80’s initiative called Just-In-Time manufacturing. This is sort of an advanced JIT.

A: Yes, absolutely! Totally correct! Threadless and mass customization companies have found a great way to reduce waste and access inventory by tapping into the preferences of their potential customers before they invest in material production. And once they know what customers exactly want, they deliver fast!

Q: This leads me to a question of suppliers and their relationships. Have you seen a change in business-to-business (B2B) as a result of crowdsourcing?

A: Hmmm…this is a very good question. At the moment crowdsourcing is discussed mostly in consumer markets. But there are also examples from B2B. One example where crowdsourcing is very successful in a B2B is Innocentive, an intermediary that enables a new way to perform scientific research in the pharmaceutical industry or other chemical industries. They crowdsource technical problem solving. Companies like Eli Lilly or a Pfizer would typically solve the chemistry problems in their own labs or assign it to an external research lab or university lab. In this model, with the help of Innocentive, they crowdsource their problems to a huge open community of hundreds of thousands of scientists who then self-select if they pick a problem and work on it. This is a really appealing case of crowdsourcing used in a very professional B2B situation .

Karim Lakhani, a former Ph.D. student at MIT and now a professor at Harvard Business school, studied this crowdsourcing approach at Innocentive. He found that problems posted there were problems which the seeker companies internally tried to solve before. These are large corporations with big R&D labs. Still, they could not solve these problems after trying between six months or even two years. After the same problem were posted on Innocentive, an external solver found a solution in average after only 74 hours, compared to six months to two years before. This works as solvers already knew the solution. This is one of those really strong cases of the high efficiency of crowdsourcing. It helps you to tap into really new knowledge sources.

Q: So we really are smarter together, huh?

A: Yeah, and we’re especially smart if we ask people we don’t know. But the problem is how can we ask someone we don’t know. This is one of the largest problems in innovation management, called the local search bias. You’re only doing normally a local search for solutions within your local knowledge base. So one of the major tasks for innovation is that you combine different knowledge sources. I think this is a fundamental mechanism of what crowdsourcing allows if you don’t define who can contribute, but you leave this open and you can tap into new knowledge. This is one of the fundamental effects and why it is efficient.

Q: It would seem that it would be critical to the efficiency of the crowdsourcing model that the supplier relationships are capable of responding and delivering in flexible and timely ways.

A: Yes, in Mass Customization this one of the things that in some companies they customize according to bundled, standardized components.

Q: Like Dell?

A: Exactly. With Dell, the suppliers don’t have to be involved in the customization. If you take the European or Japanese automotive industries where you really configure each car for the consumer, each car is made on demand, their suppliers are closely integrated in the factory. Each seat is made just in time for a particular order according to a consumer’s preferences.

Q: What do you see as a roadblock to crowdsourcing? It all sounds interesting, but ….?

A: The largest hurdle is in change management. Using crowdsourcing in a corporate context is a new kind of doing business and you really need the commitment of the people in the company. In mass customization we found that many companies are good at setting up a MC pilot but really bad at turning it into a business model that is sustainable. Take Levi’s - they are still held up as an example of mass customization, but after 10 years of customizing jeans they closed their custom operations (in 2003). The main reason for this was that after 10 years they couldn’t change their mass customization pilot into a new business model. No pilot survives over 10 years. They did not realize that they could enter the kind of learning relationships with their consumers. After 10 years you couldn’t reorder a pair of custom jeans based on the information they had on you! It was not in the thinking of Levi’s.

It is the same with customer integration in the innovation process. Many companies are getting better and better in setting up the toolkits and methods for innovation, having Web sites where users can submit ideas, etc. But all of these are just extra costs if they don’t have a strategy and dedicated processes to use this external input and include it in their internal operations. In innovation management this problem is called NIH - “not invented here.” It's the resistance of insiders to accepting knowledge from the outside. In regard to crowdsourcing, this is the largest hurdle in many companies, as these ideas are not processed internally. German car manufacturer BMW has a tool on their Web site where they get over 1,000 different technical ideas from outsiders. But from those ideas, only a very small fraction is used internally, and all the others die. This is one the major risks in the crowdsourcing movement within a corporate context – that managers get excited and set up these systems but then their colleagues internally don’t process the external input and say “What is this, what’s up?” and then people will build resistance to it.

Q: In an earlier interview you talked about Proctor & Gamble’s theme of “proudly developed elsewhere.”

A: Yes. I think this is a very cool term from them that expresses the importance for companies to change their internal mindset. This is, at the moment, what I would say is most "in the way" of crowdsourcing - that most managers think in old ways of internal value equations.

Q: It is a classic complaint that I hear in my work that the employees feel that no one ever listens to their ideas. They stick out suggestion boxes and went to focus groups and raised their hand, they offered ideas and nothing ever happened.

A: Yes. We already have the problem that within a company managers don’t listen to their own employees, then why should they listen to someone from the outside that is already far away.

Q: It seems that crowdsourcing is a great way to get at new ideas and a fast way to question basic assumptions around the basic business models, but does anyone really to want hear it?

A: It may be a fault of us as management professors. We mainly teach in the manner of Porter’s Value Chain. In Porter’s value chain concept, however, there is no mention of customers as value creators anywhere. It has an internal focus. In economics we still teach that either you can provide a task internally in the hierarchy or that you outsource it to a defined provider. Crowdsourcing is a third way to organize the share of labor. You openly ask and everyone can contribute and self-select how much effort they invest in solving the task. It is really different from the economic behavior we teach our students in how a big task in a company is performed.

Q: So this is a paradigm shift that is happening as a result of crowdsourcing?

A: Absolutely a paradigm shift. In many cases it is happening anyway and companies cannot do anything. Take Open Source. Contributors to Open Source software projects don’t really care if the company considers their input or not. They just have a specific need and they organize themselves to get the job done. The beautiful thing with crowdsourcing and the Internet is that we don’t really need the infrastructure of a company. We now enter a world where the infrastructure is easily accessible to anyone who is interested in something. Then a lot can change. This is really exciting - one the most exciting aspects of this paradigm of crowdsourcing is that we don’t need companies any longer in the way we did.

Q: Wow – that could be a scary thought.

A: Being a consumer today is very exciting. Consider emachineshop.com. They are like a Kinko for machinery. You download a very easy to operate free CAD software. Then you can design what you want, upload 3D designs from the Internet, and then you can place the design on a huge park of machines - drilling, laser lathes, CNC cutters, whatever you need. You can select materials and you push a button and start a remote production process. Two days later, your design arrives at your home. So you have an entire machine park without any of the transaction costs. It is really incredible what you can do there.

My favorite example is IKEA. You find many IKEA furniture designs today reverse-engineered by users somewhere on the Internet. Say, you find a specific table you like, but you want a modification. Normally there is no way to get your own design from IKEA. But now you just download the 3D design of this table from the Internet, you make the modifications, upload it to emachineshop or a similar provider, place the order, and you get your design. This is possible today, and the premium is less than one would expect - as you as the consumer do all the design work.

Q: So the paradigm or basic assumption we are challenging is, do we really need companies at all?!

A: Yes. For most things we are satisfied with the standard model. But some products where we care or have a specific need, we will get involved. We still need the traditional model, so I don’t think that the high variety production we have today will ever die, but with crowdsourcing we have an additional model that will be utilized in the domains where they care about it.

Q: So the real question is, what is the new function of a company with the inclusion of this third model?

A: Yes, and as a company, what is really the core of a company? I challenge my students with this question when we talk about crowdsourcing. Take the IKEA case. If you get the designs and you do the modifications on the Internet, then what is the core of IKEA? Branding? If you take the Threadless guys, their core competency is building and managing the community of contributors. This is a difficult core competency and very different from traditional companies.

Q: So then these transitions from crowdsourcing are what companies must embrace before utilizing the model?

A: Yes, if they want to use it as a sustainable practice. Some of the crowdsourcing ideas you can pilot easily. If you want to build it into your business model, you have to have a lot of change in the company.

Q: What would you say has been the biggest surprise or learning for you in all of this work?

A: Well, one of these was really these Threadless founders. We invited them twice to MIT and when I had them here the first time, these were 23-year old guys that couldn’t talk in public and they were dressed like freaks, but already at this time were making half a million dollars profit each month without any risk investment, employees and capital. That was really the surprise of it all. I invited them to a company group of, like, 20 managers and we all saw these guys and said, "Wow, this really is a new world." If you see these guys – and they were telling us they do, like, five hours of work and 10 hours of party each day - they were really totally different to the Silicon Valley entrepreneurs. You always knew there were young people with cool ideas who made a lot of money on the Internet. They were so different.

Q: That’s the one that stands out the most.

A: Then there is the very, very beginning, a long time ago, it was mass customization in ’94 in New York City, and we saw this first Levi’s custom jeans operations. This was before the Internet and everything and they were customizing jeans in New York City for a very very small premium and I just had read about it before in an article, and then I saw it. I said, "Wow, the reality was there." This is probably how I started, I thought it would be a nice topic for a Ph.D..

Q: So, the mass customization idea was already in play before Pine and Gilmore talked about it?

A: Well, yes, the mass customization idea was probably first really described by Alvin Toffler, who did a book, “Future Shock,” in 1970 where he described mass customization without using the term, and then the term - it was in connection with his computer integrated manufacturing, flexible manufacturing - so, he was thinking about what will really change now with these new kind of flexible manufacturing techniques that were coming up at that time. And he really described mass customization in a very precise way. And then the term was coined in 1989 by Stan Davis in a book. So he invented the term, but then Joe Pine sort of made the concept his master’s thesis on mass customization at MIT in a master’s program here.

Q: That was in the ‘90s.

A: That was ’91. And then he made out of his master’s thesis his best selling book “Mass Customization” in ’93. And this is how I got interested in it.

Q: So, how does this relate to your earlier work with mass customization? It appears that this way of users manufacturing their own goods is a new level of mass customization.

A: Absolutly! In the beginning, mass customization was very much production driven. Then, it was connected with the Internet. Today, we have a third generation of mass customization: companies, connecting mass customization with online communities. Consider Spreadshirt, Zazzle or CafePress. These are companies that allow you to customize goods, but you can order those not just for you, but you can sell your creations to others. So these companies have combined the eBay model of very easily selling stuff over the Internet with the customization model.

Q: The result sounds like a sort of instant company …

A: Exactly. You can very easily use the infrastructure of these companies to get your ideas into reality and even make money with this. Consider Spreadshirt. They have like 250,000 shop and sell like 100,000 products a month. So most shops don’t sell any products in a month. But no one cares, as this is not connected to any physical inventory or cost. Manufacturing is centralized, and what they do is produced on demand.

Q: The Just-in-Time model.

A: Exactly. This is pure just in time.

Q: So the other side of this then is part of your more recent work on the idea of collective customer commitment?

A: Yes, this is an alternative model! Collective customer commitment is how I called the Threadless model in a joint article with Susumo Ogawa. At Threadless, they do not do on-demand manufacturing or customization, but collect the commitment of a group of customers first where customers agree on a common standard. And then they produce this standard in a larger batch equally for everyone. This is a reverse of the long tail idea (in manufacturing) and almost a re-interpretation of mass production. But both models, mass customization and collective customer commitment, are models of crowdsourcing. In mass customization, the crowd is reacting to an open call to design an individual good; in the second, the main focus of participation of crowd is the selection crowdsourcing (product management) task.

Q: Can you imagine other functionalities of a firm where crowdsourcing could start, beyond design and selection?

A: Absolutely. A good example is after-sales service and support. In the computer industry you used to have a service contract, giving you a special number and the right to call this number and request help within a defined period of time. Today, often it is much more efficient if you go to an Internet forum and post your question to an open community, and someone will help you.

So, this probably is an alternative model. You crowdsource your question. You also could offer a specific incentive to influence somehow in the community to help you faster or in a more detailed way. But the difference is that you pick the incentive and the problem solvers self-select if they will help you and to what extent.

Q: Maybe this brings back the old barter system, huh? It seems like we’re coming full circle.

A: It’s very interesting. Before we had the Industrial Revolution, there was no large division of labor. We were doing much more within one community. But then the industrial revolution brought us this huge division of labor. And, in result, the distance between producers and consumers really grew. In a local economy, there are very close relationships between the local class people, and often we had this barter trade system.

And work and life were much more intertwined. If I were a cobbler, I was a cobbler 24 hours a day. I knew my neighbors and I got a chicken because I fixed your boots. I think it is that sense of community. It seems to me that the underpinnings of the crowdsourcing is a return to a sense of community.

That’s a nice idea and the underlying of many crowdsourcing models. But also the other way around can be true: consider Amazon’s Mechanical Turk. It is used to outsource very small (boring) highly defined tasks like editing pictures, writing small pieces of code or whatever, to a large crowd of people. And you get paid one or two cents for one task. For people contributing here, this may have to do nothing with a community. But over all, I would say it’s exactly what you said before - crowdsourcing builds on these communities, thinking of people of giving and getting something back.

Q: That’s why I chose your term “learning relationship.” I think there are other forms of crowdsourcing relationship that we are just starting to understand.

A: Yes. This is probably wide open for researchers. One of the open questions maybe where we don’t know very much about is what is really the kind of relationship the company crowdsources with his customer. How are the resulting relationships different from traditional relationships? We hear some assumptions, but there is no really large peer group associate that looks in to what is really changing there on the long term.

Q: Well, very interesting. When I was putting together my questions one of the things that popped out you just brought up - the Industrial Revolution. I think in an interview with mass customization pioneer Joseph Pine on your blog, Joe noted that Henry Ford was the father of mass production, but Michael Dell is the Henry Ford of mass customization. So who’s the Henry Ford of crowdsourcing?

A: That’s a good question.

Q: Frank Piller?

A: No, no, no. I’m just writing about it. It is someone who is doing it. Wow, I think this is a question you should give back to the entire Assignment Zero team. I think you’ll probably get a top 10 list. On my top 10 the Threadless guys are on it, and so is the founder of IStockPhoto, but also Jimmy Wales, founder of Wikipedia, Linus Torvald, and so on. You should really propose it to this community to make like a top 10 of the Henry Ford's of crowdsourcing.

Q: One of the other articles that you had that I was struck by is when you were talking to Detlef Schoder, a professor from the University of Cologne, about crowdsourcing the mass customized newspaper. This really is supposed to be a standard product. Have you seen anything like this?

A: Up to now, no. This is an old idea and technically it’s pretty easy, but in a printed form, that the newspaper you get in the morning to your door is customized; no one did it yet. The closest to it is the Wall Street Journal online addition. It’s one of the most successful online products where people really pay for content. And most of the subscribers to the Wall Street Journal Online get a personalized edition. But this is a pure online product. Even the technology is there to get a custom paper delivered to your door. But would consumers accept this? This is a real challenge, not so much for producing or configuring it, but it really demands a change in our behavior and perception of the newspaper. And perhaps, we don’t really want to know what we read. We want the surprise of a newspaper.

Q: And I don’t want to only know what I already know, I want to have the feeling that I’m learning what I don’t know.

A: Exactly. And also you want to have a common thing to talk to your colleagues at the office, you know, and other common things. So this is an example of a product where I see some restrictions to customization, and also crowdsourcing in general. On the other side, this particular project that provides the background for the interview we are just in, Assignment Zero, provides a great example for a new form of journalism taking place in the form of crowdsourcing.

5/21/07